What makes a game fun? Episode 003 – Snowballing

What Makes a Game Fun 003

Now, this’s a bit off the topic I’d been going to cover, but it’s pertinent to the first two episodes of the “What Makes A Game Fun?” series. So what’s the topic that I’m beating around the bush on today? Why, snowballing.

No, no, not THAT kind of snowballing… ew. It’s also best if you don’t ask yourself why I know that term. No, I mean the kind of snowballing which refers to the effect of a snowball rolling downhill; it rapidly picks up power and progressively gets bigger and bigger.

In game design, this translates to recursive positive reinforcement.

To put that in layman’s terms, if you do well in a game, it rewards you by making you stronger, thereby making it easier to continue to do well, which rewards you again in turn, each winding growing progressively easier.

There are a lot of individuals, both players and developers, who consider snowballing to be a major problem, one which needs to be addressed. Then there are those, such as myself, who hold to the position that snowballing is both beneficial and, in many cases, an absolute necessity for the proper design of many games.

So, why would there be a debate at all? How can something be both good and bad?

Fortunately, we have a term for this: it’s called a paradox.

The-Pinnochio-Paradox

As we’ve seen in the past two editions, most of game design is comprised of balancing out mutually exclusive goals. It’s not a matter of making both sides perfect, as it’s often literally impossible to please both ends perfectly, but rather the end goal is for the two sides to be balanced with a target audience in mind, taking their personal tastes as a group into account.

In terms of snowballing, the two distinctively opposing ends of the spectrum are as follows:

Arguments for snowballing:
– The need for player progression
– Positive reinforcement of desired player interaction
– Refusal to penalize players for doing what they’re supposed to

Arguments against snowballing:
– Early advantages grow too fast causing the rest of the game to be a moot point
– Blind luck can drastically favour someone all game
– Permanent benefits for temporary actions

One will quickly note that the key in all of these is the fact that snowballing, by its very nature, tends to involve recursive positive reinforcement. Note, however, that just because it’s called positive reinforcement doesn’t mean it’s a good thing, nor that you’re reinforcing what you want to.

positive reinforcementTo make this a bit more clear, let’s take the example of a MOBA since they’re especially well known for their snowballing being taken to absurd levels:

Player 1 kills Player 2. This is good, this is what we want to have happen in the game and as such a number of things occur.

+ Player 1 is rewarded bonus gold for killing Player 2.
+ If it was first blood, also known as the first kill of the game, there may be even more gold given to Player 1.
+ Player 1 gains bonus Experience, possibly giving them a level advantage over Player 2.
+ Player 1 has some free time to do whatever they want relatively risk-free.
+ More free time means more gold and exp while safely farming.

– Player 2 is penalized gold for dying.
– Player 2 is penalized by waiting on a respawn timer.
– Player 2 is further penalized by a delay between respawning and running back.

too much of a good thing

Too much of a good thing applies to snowballing and capsaicin roughly equally. If you hit ghost pepper, you probably should’ve stopped a long, long time ago.

With this situation, Player 1 is granted several benefits in a row, rewarding them multiple times for doing a good job. The nature of this is that Player 1 may be substantially stronger than Player 2 by the time Player 2 gets back to their lane and they’re now at a significant disadvantage. This makes it much easier for Player 1 to just kill Player 2 again, which further makes it easier to do so again, and again, and again.

In this example, an entire game can be determined based on the actions at the very start of the game. Something as small as a 10% difference in random damage can determine the winner or loser of a game 20 minutes or longer before the game actually ends, making the rest of the game a waste of time to play out. This isn’t always the case, however, so players are encouraged and/or forced to continue playing even if it seems like a lost cause.

This sounds particularly bad and makes the case against snowballing to be rather strong it would seem, however I shall point out that it’s not the snowballing itself that’s the problem, but the severity of that snowballing.

It really is a case of too much of a good thing. For example, I love spicy food. Today I had ghost pepper wings. This is much the same concept of why snowballing can be bad. It’s also a prime example of why I shouldn’t be allowed to make decisions for myself.

In this example, Player 1 is granted multiple, cumulative benefits for the kill, and Player 2 is also given multiple, cumulative penalties for being killed. Even if each benefit or drawback is individually small, the cumulative benefit of each being combined together is one of a multiplicative effect making it stronger than originally intended, since it becomes nearly impossible to predict the effects of multiple multiplicative benefits in a row.

So how does this affect the value of “fun” in a game?

To steal a partial quote from Tom Cadwell of Riot Games, it adds to the anti-fun of the game more than it adds to the fun.

“Anti-fun” is a concept that’s been around for a long time, but only in the last few years has become a well known term, and it’s generally misquoted or misunderstood, where the average person believes that there should be absolutely zero anti-fun in a game. In practice, that’s not how things work: any time Player 1 kills Player 2, there’s going to be some removal of fun of the game from Player 2.

The trick is to ensure a balance that any time fun is removed from the game, more fun is added than removed in net total.

The example given originally by Cadwell was that of a mana burn effect: removing the mana from an enemy so they can’t cast any spells may make the game a little more enjoyable for the one attacking, but it’ll make the person on the receiving end feel awful. In our above example, the first player will feel glad to have gotten the kill, and the first blood bonus may make it a bit better, but the player who had their gold removed, then irritated as they sit around waiting to do anything at all again, then further annoyed as they have to run back on foot, and then annoyed again to find that the other player is substantially stronger, now means that the total net value of fun in the game has been set to a negative value.

Snowballing has the serious potential to go many times too far beyond where it should due to its recursive nature. As such, all game designers must absolutely be certain to keep a very sharp eye on recursive effects, especially when in terms of rewarding a player for doing well, or punishing one for doing poorly.

However, that doesn’t mean we should strip out snowballing entirely. In fact, the very act of introducing a snowball effect to a game’s design can create many benefits for the overall fun level for all players involved.

First off, rewarding a player who did something you wanted them to sends a very clear message to them that they did something right and that they should try to do more of whatever it was they were just doing. This is the whole “show me, don’t tell me” method of game design where the player isn’t just told they’re supposed to perform an arbitrary task, but rather they’re actively encouraged by way of the mechanics themselves.

Installing mechanics into a game’s design where the player is encouraged to play properly rather than just being told to is a great way to increase player satisfaction. Furthermore, when it’s clear and obvious of an advantage, such as being paired with audio and visual cues, such as a loud “HEADSHOT!” and cheering with particle effects playing across the screen, it really makes the fun that much greater.

And this is why beating up little kids when you're 30 is looked down upon.

And this is why beating up little kids as an adult gives less exp.

The most obvious form of snowballing is the generic “RPG”, or role playing game, mechanic of gaining experience and levels. This level increase is a hard, quantized showcase of having done something well. Kill enemies, get levels.

Most RPGs since the 1990’s have also installed two systems which limit how strong the snowballing gets, however. The first is that levels progressively cost more exp, and the second is enemies that are outside of your character’s level range are worth reduced experience, or none at all after a certain point.

The reasoning behind such is that snowballing should only reward a player for playing properly. As such, a level 100 character farming level 1 monsters isn’t really facing much of a challenge so shouldn’t be given equitable experience compared to what the same character would’ve gotten for killing a level 100 monster.

Virtually every game these days involves some form of snowballing, with few exceptions. If there’s an economy in the game? Snowballing via money. If there’s character stats with any kind of growth potential at all? Snowballing. Any RPG elements at all such as exp and leveling or purchasing items? Snowballing again.

To get back to snowballing, the fact of the matter is that good game design will reward a player when they do something well, with a hook to encourage them to do it again, such as cumulative benefits for repeating the same effect.

Many games, MOBAs in particular, now also offer permanent progression beyond the individual game. Games such as Evolve or Hearthstone also reward players with greater strength in a snowballing manner where the more games you win, the more likely you are to win more games against people who haven’t won as many in the past.

One will note that Hearthstone is not particularly well designed (at least in terms of gameplay; it’s tuned more towards catering to the addictive personality, much in the way candy crush is – good game design actually decreases the sales potential oddly enough; we’ll get into that in another episode though) as it simply rewards a player consistently for performing the same actions but doesn’t make their task more difficult or put in any sort of counter-balance play. A player who has been building the most powerful deck imaginable can, and will, still play completely new players to the game with no capacity to fight back. In this case, as long as a player keeps winning, they’ll keep being rewarded for it without an opposing force to balance it out.

Evolve, in contrast, gives benefits for players who win a match. If, for example, the monster defeats the hunters one mission, the monster may spawn stronger minions. If the hunters had defeated the monster, however, their base may be given more powerful turrets for defense instead. The key here, however, is that if one team is consistently losing repeatedly, they’re granted bonuses to keep the match somewhat fair as a PvP game which is completely one-sided isn’t much fun for anyone. It’s mildly amusing to the winning team, but not nearly as much as if they’d beaten a real challenge. As such, Evolve is a much better example of how to implement snowballing in an interesting and effective manner.

Another good example of snowballing is Tower Wars, wherein the players are encouraged to be highly aggressive, rewarding them with more points to spend on upgrades the more aggressive they are. At the same time, players are also rewarded for killing enemy units. This causes the oddly confusing situation where the first player may send a unit to attack and is given points to make that unit stronger, but the opposing player is given a gold bonus for killing it, making it more likely to kill the next one. This means both players are snowballing in different ways off of the same action so that, instead of counteracting the snowball of one player, both players are snowballed for performing their opposing tasks properly and it evens out remarkably well.

So there are some distinctive advantages to including snowballing into the design of a game, yet many designers ask what to do about the problem of snowballing. What they’re really asking, whether they know it or not, is what to do about the problem of adding “too much” snowballing, not snowballing as a whole.

The correct answer is to simply limit how often you apply it so it doesn’t get out of hand in the first place. Trying to counteract the very nature of snowballing causes a very real problem: actively preventing snowballing means introducing elements such as rubberbanding or handicaps. These, by their nature, cause the problem of penalizing players for doing well, and rewarding players for doing poorly. Yes, I just congratulated Evolve for doing exactly that, yet the term “all things in moderation” is a phrase to live by as even supposedly “bad” things can be good in small amounts.

Well, more like a little bit of a bad thing can be good. Close enough.

Well, more like a little bit of a bad thing can be good. Close enough.

The point is, however, that if you want someone to do something, you reward them for doing it. If you want them to not do that anymore, you penalize them for it. This isn’t a complex concept, and yet it’s something many game designers have severe issues with when it comes to trying to make a game balanced and fair.

As such, to be perfectly blunt and clear for any who missed it: if there is a snowball mechanic in a given game, it should be there because the player is meant to be rewarded for doing well. If snowballing is far too strong, trying to counteract the mechanic tends to backfire on the whole purpose of the snowballing mechanic in the first place as the only real way to accomplish it is to penalize that same player for doing exactly what you wanted them to do, thereby sending conflicting messages to the player as to whether it was good or bad to do something.

Once conflicting messages are applied, this also creates the effect where players may very well intentionally perform poorly so as to gain the rewards of doing poorly, then using those rewards to do well. Or worse yet, simply not caring if they do well or not since they’ll be rewarded either way.

Unlike what some individuals or groups may claim, telling everyone they’re a special snowflake and get an A for effort doesn’t really work in practice, especially not in the gaming community. Gamers want to be rewarded for doing well, not for simply showing up.

Back on topic, the final major benefit about snowballing is that many games are very much a tug-of-war of sorts, where two opposing players or teams are trying to win. In some games, this can lead to unwinnable games which have ground to a stalemate where neither side is able to effectively combat the other, or where games drag on for inordinate amounts of time without a clear victor. Snowballing, in these cases, can be used effectively to give the team which is doing better a strong enough advantage to push for a win and let both teams get on to their next game quicker.

To sum it up, how you view snowballing is largely dependent upon the perspective you’re looking at the issue from.

For those who create games: only add snowballing mechanics as needed, and don’t overuse them. Counteracting snowball mechanics is invariably a bad design decision in most cases, so if you don’t want players to be rewarded for doing well, don’t reward them in the first place. If you do want them rewarded, be cautious to not go full monty haul: don’t fork over huge bonuses for simple tasks as it’ll quickly grow out of control.

For those who review games: look specifically at the game’s mechanics to see whether it rewards the player for doing what they’re supposed to or penalizes them for it. Also check if a player is rewarded or penalized multiple times for the same action; for a good rule of thumb, if a player is rewarded more than twice for the same action, it’s probably too much and should be reflected in your review.

For those who play games: keep in mind that 99% of game designers aren’t trying to be jerks and they really don’t have it out for you. Snowballing is added to a game specifically because it makes you happy when you get rewarded for doing well – it’s part of our innate psychological makeup. When snowballing goes too far by making the game trivial in difficulty, or is counteracted by penalizing players for doing well, then the game itself is suffering severe issues and you should voice your concerns to the dev team on their forums, in tweets, or however possible.

Remember, the designers are trying to make a fun game by juggling mutually exclusive goals, such as rewarding players for doing well while still keeping it balanced. Keep your criticisms focused on where this is breaking down and why it’s an issue. A good designer will listen to their critics, but may not directly respond to them. They’re listening to what you have to say, though, so make sure you try to provide as much information as to what the problem is so they can fix it. If you’re upset, wait until you’re calmed down some, then try to explain your point of contention clearly.

Now go forth and find a game that involves snowballing and play it again. Think about the differences in it now that you have a new understanding of the issues, and see if you can spot why each design decision was made – you may be surprised at what you find and that the bits that frustrate you sometimes may very well have a good reason for their inclusion!

About Catreece MacLeod

Catreece MacLeod has worked as a writer, editor, video game designer, teacher, 3D artist and quite a few other roles. Her specialization is pre-production writing, most notably, world design and IP creation.
Bookmark the permalink.